Open phamthanhdat270198 opened 2 months ago
You need to adapt the inference code, create multiple streams and pass to infer() function.
When I create multiple streams It just decrease a little bit infer time. So Does it working in parallel. give me an example about multi streams and multi contexts. Many thanks!!! Here is my impliment multi streams ![Uploading image.png…]()
Env
Question
Hi, I want to set specific cudastream ID on Cpp code yolo like --streams=0 or --streams=1, ... in trtexec Many thanks!!!!!